Variability in ice motion at a land-terminating Greenlandic outlet glacier: the role of channelized and distributed drainage systems

Abstract

ABSTRACT. We use a combination of field observations and hydrological modelling to examine the mechanisms through which variability in meltwater input affects ice motion at a land-terminating Greenlandic outlet glacier. We find a close agreement between horizontal ice velocity, vertical ice velocity and modelled subglacial water pressure over the course of a melt season. On this basis, we argue that variation in horizontal and vertical ice velocity primarily reflects the displacement of basal ice during periods of cavity expansion and contraction, a process itself driven by fluctuations in basal water pressure originating in subglacial channels. This process is not captured by traditional sliding laws linking water pressure and basal velocity, which may hinder the simulation of realistic diurnal to seasonal variability in ice velocity in coupled models of glacial hydrology and dynamics.

Dynamic Response of a High Arctic Glacier to Melt and Runoff Variations

“…with the highest mean correlation for the two regressions. Based on previous work, we argue that the decline of sensitivity of velocity to runoff around this threshold is likely to arise from a change in drainage morphology from a low-efficiency distributed drainage system to a highly efficient channelized system (e.g., Bartholomaus et al, 2008;Cowton et al, 2016;Iken, 1981;Kamb et al, 1994). Slopes of the linear regressions ( ) represent the sensitivity of flow velocity to runoff during the early melt season, that is, prior to channelization, and during the late melt season, that is, after channelization ( Figure 3).…”

“…The local cumulative runoff for each velocity anomaly is indicated by the marker color. Based on previous work, we argue that the decline of sensitivity of velocity to runoff around this threshold is likely to arise from a change in drainage morphology from a low-efficiency distributed drainage system to a highly efficient channelized system (e.g., Bartholomaus et al, 2008;Cowton et al, 2016;Iken, 1981;Kamb et al, 1994). To determine the cumulative runoff at which this transition occurs, two linear regressions were performed: one for data above and one for data below a cumulative threshold value.…”

“…), which has previously also been found by Bartholomew et al (2010) and was ascribed to short-term (daily) peaks in runoff, causing short-lived pressurization of the channelized and adjacent distributed system. It has been argued that steady state theory, suggesting an inverse relation between runoff and water pressure during channelized drainage, is unlikely to apply in environments with rapid variations in runoff (Bartholomew et al, 2012;Cowton et al, 2016;Hewitt, 2013;Kehrl et al, 2015). It has been argued that steady state theory, suggesting an inverse relation between runoff and water pressure during channelized drainage, is unlikely to apply in environments with rapid variations in runoff (Bartholomew et al, 2012;Cowton et al, 2016;Hewitt, 2013;Kehrl et al, 2015).…”

“…during the melt season, the drainage system expands, and when discharge exceeds a critical threshold (Kamb, 1987;Walder, 1986), the drainage system transforms into a more efficient channelized drainage system (e.g., Bingham et al, 2005;Hock & Hooke, 1993;Nienow et al, 1998;Röthlisberger, 1972), operating at low water pressure and draining water from the connected distributed drainage system (Andrews et al, 2014;Cowton et al, 2016;Hubbard et al, 1995).…”

“…Fast ice flow during channelized drainage may occur in response to high short-term runoff variability, since cavity expansion in the connected distributed system adds a forward component to ice flow, while cavity closure is more downward (Cowton et al, 2016;Iken, 1981). It has been argued that steady state theory, suggesting an inverse relation between runoff and water pressure during channelized drainage, is unlikely to apply in environments with rapid variations in runoff (Bartholomew et al, 2012;Cowton et al, 2016;Hewitt, 2013;Kehrl et al, 2015). Since we find a strong positive correlation of 10-day runoff and the variance of daily runoff within 10-day intervals (R = 0.79; P < 0.001), we expect that the positive relation between runoff and velocity at a 10-day time scale may well be the result of runoff fluctuations at time scales shorter than 10 days.…”